A PLACE on the FRINGE of SAGALASSOS the Excavations at the Rock Sanctuary

A PLACE ON THE FRINGE OF SAGALASSOS The excavations at the Rock Sanctuary

Peter TALLOEN, Philip BES, Mücella ALBAYRAK, Bea DE CUPERE, Katrien VAN DE VIJVER and Jeroen POBLOME*

Abstract

The so-called Rock Sanctuary, a distinctive limestone rock outcrop with natural cavities situated in the periphery of the Pisidian city of Sagalassos (SW-Turkey), was a natural feature that was served a variety of functions throughout its history. Rescue excavations carried out at the site mainly yielded evidence for the deposition of specialised offerings in the form of ceramic, glass, metal and stone vessels, pieces of personal adornment, instruments for textile production, but especially many thousands of fragments of terracotta figurines. All of these identified RS as a ‘special-purpose site’, a natural landform that was given a cultural significance, not by means of monumentalisation but through the activities that took place there during the Hellenistic and Roman Imperial periods. It was the combination of all these objects as a whole and the very context in which these were used and placed that made it possible to identify the site as a sanctuary, more particularly, a site of popular worship. This paper presents an overview of those excavations, highlighting the significance of this site in the landscape of Sagalassos and what it can tell us about the community that conceived it and used it as a cult site, outside of the sphere of official religious practice. RS thus offered a unique glimpse into an aspect of ancient life not previously known from Sagalassos.

* Peter Talloen is Assistant Professor of Archaeology at the Süleyman Demirel University of Isparta (SDU; Turkey) and Assistant Director of the Sagalassos Archaeological Research Project (SARP); Philip Bes is an independent researcher from the Netherlands and a Roman pottery specialist of SARP; Mücella Albayrak is a PhD candidate in Archaeology at SDU; Bea De Cupere is an archaeozoologist at the Royal Belgian Institute of Natural Sciences (RBINS); Katrien Van de Vijver is an archaeo-anthropologist at RBINS; Jeroen Poblome is Full Professor of Archaeology at KU Leuven (Belgium) and Director of SARP. The study of the Rock Sanc- tuary was supported by the Belgian Science Policy Office, the Research Fund of KU Leuven and the Research Foundation Flanders. The authors would like to thank the Ministry of Culture and Tourism of the Republic of Turkey, its Kültür Varlıkları ve Müzeler Genel Müdürlüğü, and its representatives for the excavation permission, support and most appreciated aid during the 2014-2018 fieldwork campaigns. They especially wish to thank the former director of the Burdur Museum, Hacı Ali Ekinci, for his help in facilitating the rescue excavations. Finally, other scholars who contributed to the study of the different find categories excavated at RS are acknowledged: Dries Daems (Hellenistic pottery), Veerle Lauwers (glass), Ralf Vandam (prehistoric pottery) and Rinse Willet (cookware).

INTRODUCTION

During the study campaign of June 2014 a collection of terracotta figurines, stored in the depots of the Sagalassos Project in the village of Ağlasun (province of Burdur, Turkey), were reviewed for publication. This collection included a group of mould-made terracotta figurines dating to the Roman Imperial period, which had been confiscated in 1991 by local authorities and handed over to the Sagalassos Project for safe storage in the depots of the excavation house. The figurines allegedly originated from illegal excavations in the vicinity of the archaeological site of Sagalassos. Upon inquiry as to the exact origin of the artefacts, team members of the Sagalassos Project were taken by a local shepherd to a location some 600 m to the southeast of the urban centre, immediately beside the modern road leading to the ancient city (Fig. 1).

Fig. 1. Location and aerial image of the Rock Sanctuary (Google Earth 2017).

This proved to be a limestone rock outcrop in the mountain slope with several large crevices covered by limestone boulders, thus creating several cavities within the mass of the rock (Fig. 2). The abundant presence of artefacts on the surface both within the crevices as in its immediate surroundings in the form of many hundreds of fragments of locally produced terracotta figurines, but also of ceramic and glass vessels dating to the Late Hel- lenistic and Roman Imperial periods, confirmed this site as the source of the confiscated terracotta figurines. Furthermore, the exceptional nature of these finds in terms of type and quantity allowed this location to be identified as a site of ‘specialised deposit’” of artefacts1. While terracotta figurines could be deposited in a number of contexts, such as

1 Alcock and Rempel 2006. A PLACE ON THE FRINGE OF SAGALASSOS 251

Fig. 2. View of the rock outcrop from the northwest (© Sagalassos Archaeological Research Project).

domestic contexts and burials2, their exceptional quantity in a cave-like location, which was often considered numinous in antiquity3, pointed to a cultic site. This was an area set apart for the worship of gods, a holy place where people went to undertake religious rituals in the form of sacrifice, prayer, and the giving of votive offerings. Accordingly, the site was designated as the ‘Rock Sanctuary’ (hereafter RS). Dumps of excavated material present throughout the site indicated that it had fallen victim to many years of illicit digging, which destroyed much of the stratigraphical record. This undoubtedly caused the loss of numerous finds, which were sold to private collectors and museums all over the world. Most of those pieces most probably ended up in Turkey: at the regional archaeological museums of Burdur and Isparta 330 of the 541 registered figurines (or 61%) and 70 of the 140 registered figurines (or 50%) respectively could be generally attributed to the production at Sagalassos on the basis of typology and clay fabric, and in some cases actual fits between fragments kept at the Burdur Museum and fragments excavated at RS leave no doubt that these were brought from RS4. Also, several pieces kept at the Sadberk Hanım Museum in Istanbul most probably originated from the site5. More exceptional figurines in terms of preservation and suspected to have come from RS have also been recognised among the collections of museums outside of Turkey such as the National Museum of Antiquities in Leiden, the Princeton University Art Museum, and the Martin von Wagner Museum in Würzburg, and among items that were sold at major auctioneers such as Christies’ and Sotheby’s6. Given the fact that the site is easily accessible for clandestine activities, as attested by the illegal excavations, permission was sought for immediate scientific investigation of the sanctuary in the form of rescue excavations in collaboration with the directorate of the

2 Huysecom-Haxhi and Muller 2015. 3 Mavridis et al. 2013. 4 Talloen 2020. 5 Talloen and Özden-Gerçeker 2020. 6 Talloen and Özden-Gerçeker 2020. 252 P. TALLOEN ET AL.

Fig. 3. Plan of RS with indication of the different rooms (measured and drawn by Ö. Başağaç; © Sagalassos Archaeological Research Project).

Archaeological Museum of Burdur. Firstly, the aim of these excavations was to clear completely the deposits of previously excavated soil and salvage whatever artefacts remained in them. Furthermore, the spaces inside the rock outcrop were investigated in the hope of finding in situ deposits not touched by past illicit excavations, which could inform on the occupational history of the site and provide a more precise chronology for the different categories of material culture that were deposited there. In view of its potential for understanding ancient religious practice in the region, a second aim of the excavations concerned the nature of the cult or cults that were practiced at RS, i.e. which gods were worshipped, by whom and by means of which religious practices. These questions would be answered through an examination of the situation and physical form of the sanctuary on the one hand, and the residue of ritual activity, in the shape of diverse classes of material evidence, embedded in it on the other. In order to achieve these aims a series of trenches was planned in the three areas encom- passing four spaces – designated Rooms 1 to 4 – of the cave-like complex that had not completely collapsed and therefore were (partly) accessible for research (Fig. 3). These areas were investigated over four excavation campaigns at the site, between 2014 and 20187. First, these spaces and their stratigraphy will be addressed, before we turn to the occupational history of the site.

7 Talloen et al. 2015; Poblome et al. 2019. A PLACE ON THE FRINGE OF SAGALASSOS 253

SPACES AND STRATIGRAPHY

As mentioned above, RS is not an actual cave site but consists of a rock outcrop with several large rock crevices that were covered by huge limestone boulders creating a number of cave-like spaces. On the outside no obvious remains of any man-made structures or traces of stone carving could be observed, and also inside the natural appearance of the outcrop was largely left unchanged. Our current understanding of RS is severely hindered by the collapse of the ceiling of those spaces, making large parts inaccessible and the spatial analysis of the complex a difficult enterprise. Three main working areas were distinguished: a ‘South Zone’ with the probable entrance to the complex and Room 1 in the southeastern part of the outcrop, a ‘North Zone’ at the northwestern end of the sanctuary complex, that encompassed Room 2 as well as an open-air area where numerous traces of illegal excavation were present, and a ‘Middle Zone’ at the centre of the rock outcrop comprising Rooms 3 and 4.

South Zone The cave-like complex was originally accessible from the southeast through a 1.05 m wide opening between the faces of the outcrop and covered by a lintel-like boulder which has now collapsed (Fig. 4). The badly weathered rock faces on either side of the opening display possible traces of ancient stone carving to facilitate access to the southernmost room of the complex, Room 1. This roughly rectangular ‘room’ with a southeast-northwest orientation is a covered, corridor-like space with a length of 4.15 m and an average width of 1.20 m. Its walls were formed by the faces of the limestone bedrock and its ceiling consisted of large limestone boulders which were lying on top of the bedrock, thus form- ing a covered crevice. At the northern end of the space an extension was present on both the east and the west side; the latter gave access to a further space in the central part of the outcrop (Room 4) which has now been blocked by the collapsed ceiling of both spaces.

Fig. 4. View of the collapsed entrance to Room 1 in the South Zone (© Sagalassos Archaeological Research Project). 254 P. TALLOEN ET AL.

The surface of Room 1 was littered with ceramics, including pottery roughly dating between 200 BCE and 200 CE, as well as few Byzantine shards. Large amounts of fragmentary terracotta figurines were retrieved, mainly representing different types of the goddess Aphrodite (such as anadyomene, genetrix, Knidian, pudica and strophion). Other deities included Athena, Eros, Hermes, Kybele and Tyche, while numerous fragments of female busts were also present. Further finds included shards of glass vessels, animal bones, and a seashell. Underneath, we exposed a 0.16-0.60 m thick deposit of previously excavated material, again with many fragments of pottery and terracotta figurines dating to the same periods as the material found on the surface. This deposit was sloping down towards the south into Room 1 from an adjacent space (Room 3) in the central part of the complex, indicating that the latter space was the origin place of deposit of this material, which had been redeposited as a result of illicit excavations. The excavated soil was present on top of a 0.20 m thick deposit of colluvium coming from the same space to the north. After removing the deposits of illicitly excavated soil and the underlying natural colluvium, a sequence of in situ deposits was recorded in this part of the complex. The substrate of the uppermost ancient walking level was a 0.25 m thick deposit of compact sandy silt. Its ceramic content was mixed, including pottery ranging between 200 BCE and 200 CE, and some Middle Byzantine shards which provided a terminus post quem for its lay-out. At the northern extremity of Room 1 the 0.50 m thick fill of a 0.52 m wide (NW-SE) and 0.62 m long (SW-NE) oval pit was excavated. It included shards of Middle Byzantine incised sgraffito and champlevé pottery which could be dated to the late 12th-early 13th century8, bones of sheep/goat, chukar (Alectoris chukar) and hare (Lepus europaeus), as well as ashes and charcoal, which suggest some fire-related activity, possibly cooking, in the space. Underneath, a layer of colluvium with a thickness varying between 0.17 m and 0.40 m, and consisting of dark brown silty soil with small limestone fragments contained some ceramics from the Roman Imperial period. It had accumulated over several large limestone fragments, most probably part of the collapsed ceiling which had fallen, possibly as the result of an earthquake, on top of a floor level (Fig. 5). The associated floor substrate with a thickness of 0.10 m could only be given a post quem date of the Roman Imperial period on the basis of the pottery it contained. It was arranged on top of two levelling deposits of silty soil with large amounts of limestone fragments and a thickness of 0.13 m and 0.22 m respectively, containing ceramics dating to the 1st-early 2nd century CE. Due to the presence of the large limestone chunks the excavations could only continue in the northwestern part of Room 1, over a length (E-W) of 2.25 m and with a maximum width (N-S) of 1.00 m. Two lower floor substrates of dark brown silty soil with many limestone and charcoal fragments and with a respective thickness of 0.16-0.29 m and 0.35- 0.46 m were encountered. The ceramics they contained provided a terminus post quem in the 1st century CE for the arrangement of these deposits. Residual Hellenistic pottery was also present in considerable quantities in these two floor substrates, but unlike the

8 We are grateful to Athanasios Vionis (University of Cyprus, Nicosia) for sharing his expertise with us through photographs of these finds. A PLACE ON THE FRINGE OF SAGALASSOS 255 deposits in the other spaces of the complex (see below), only few fragments of terracotta figurines could be retrieved from these layers; furthermore, they included two iron writing utensils or styli. The two floor substrates were laid out on top of a series of fills with a combined thickness of 1.25 m consisting of brown silty soil and small and medium-sized limestone fragments, placed between the two converging faces of the rock outcrop which formed the northeastern and southwestern border of the space. These fills were most probably arranged to create a horizontal walking level within the crevice. Remarkably, these fills, which constituted the lowest recorded deposits in Room 1, only contained prehistoric ceramic material, dating to the 4th millennium BCE and suggesting an occupation of the crevice during the Late Chalcolithic period.

North Zone The working area of the North Zone, situated at the northwestern end of the site, encompassed a second space within the rock outcrop, designated Room 2, as well as the area outside this covered crevice, immediately to the north. The area outside Room 2, formed by the sloping surface of the rock outcrop and measuring 5.90 m (E-W) by 2.75 m (N-S), was covered by two deposits (with a total thickness of 0.70 m) constituting the dump from illegal excavations as indicated by the presence of batteries, cigarette butts and wrapping material within the soil (Fig. 6). This material appeared to have been removed from inside the ‘room’. The layers of excavated soil dumped in the open area northwest of Room 2 contained large amounts of pottery shards ranging in date from the Hellenistic (2nd century BCE) until the Middle Byzantine period (10th-12th century CE) with most pottery dating between

Fig. 5. Limestone fragments Fig. 6. View of the area of the collapsed ceiling of Room 1 outside Room 2 from the southwest (© Sagalassos Archaeological Research Project). (© Sagalassos Archaeological Research Project). 256 P. TALLOEN ET AL. the 2nd century BCE and the early 3rd century CE. Interestingly, again some shards of Late Chalcolithic pottery, attributed to the 4th millennium BCE, were present. Other finds from these deposits included pieces of hairpins and a distaff in worked bone, miniature lead vessels, copper-alloy pieces of jewellery such as necklaces and earrings, and shards of glass and stone vessels. A Late Hellenistic bronze city coin of Sagalassos depicting the head of Zeus on the obverse and two confronting goats on the reverse was also found. Especially fragments of terracotta figurines dating to the Roman Imperial period were again prominently present in these deposits. As in the other areas of the site, the assemblage of figurines was dominated by the goddess Aphrodite and her son Eros. Other identified deities included Tyche, Hermes, Athena, Nemesis, Kybele, Isis and the moon-god Men. Frequently represented types of ordinary humans included numerous busts of women, as well as some busts of male youths. Fragments of female musicians playing harp, lyre or cymbals, and studying children with writing tablets on their laps were also present. In addition to the abundant terracotta figurines, also a single lead figurine of Hermes was found. Other than those of wild animals, such as fox (Vulpes vulpes), badger (Meles meles), tortoise (Testudo graeca) and mole rat (Spalax leucodon) which must have lived at the site after its abandonment and can also be held responsible for some of the disturbance in the stratigraphy of the site, faunal remains from this debris included chicken, sheep/goat, pig, and some little tunny (Euthynnus alletteratus) which represent consumption refuse. Also several shells of marine molluscs were present, including scallop (Pectinidae), cockle (Cer- astoderma glaucum.), clam (Glycymeris glycymeris), cowrie (Cypraea sp.) and cerith (Cerith- ium sp.). These shells are commonly found in the Mediterranean and appear to have been brought to RS as gifts or ornaments. Remarkably, the deposits also contained several human teeth as well as other, very fragmented elements of human skeletons, suggesting that the cavities of the rock outcrop were also used for burial purposes at some point. Unfortunately, since none of these were found in situ, it was not possible to determine their date. Room 2 inside the outcrop consisted of a series of interconnecting natural cavities of which the ceiling composed of large limestone blocks had largely collapsed. At the centre was a roughly triangular space with an apex oriented to the southeast, sides of 2.00 m to 2.20 m, and a base with a length of ca. 1.00 m. Additional spaces were present on the southeast and northwest sides. The SW Area of Room 2 is a roughly trapezoidal natural cavity with a length of 1.20 m (SE) to 2.80 m (NW) and a width of 1.50 m (SW) to 2.25 m (NE). The NE Area of Room 2 was a roughly rectangular natural cavity, about 1.20 m wide (E-W) and 1.40 m long (N-S). As Room 2 did not display any signs of stone carving, it was probably the result of natural erosion and karstic processes. It was also not accessible through any purpose-made entrance; a gap between the covering boulders leading to the central cavity appears to have served that purpose. These cavities were undoubtedly the origin of much of the illegally excavated material found outside the crevice, as is corroborated by fitting pieces of pottery and figurines from deposits inside and outside of Room 2. Yet, this part of the site was also characterised by a severely disturbed stratigraphy inside of Room 2, with previously excavated soil that remained present between the blocks of the collapsed cave-like space, as well as a glass beer bottle (dated 13/11/2011), a D-type alka- A PLACE ON THE FRINGE OF SAGALASSOS 257 line battery, a work glove and a shovel, most probably remnants of the illegal excavations that took place at the site. Several deposits of excavated soil with a total thickness of 0.45 m had to be removed in different areas of the SW triangular space of Room 2. From them, mixed archaeological material was retrieved. The pottery ranged in date between the Hellenistic and Middle Byzantine periods, with again several Late Chalcolithic shards present, including some shards belonging to a single vessel. Moreover, numerous fragments of terracotta figurines were found again, representing the same types mentioned above, though one figurine of Hermes was particularly well preserved. Other finds included pieces of a comb and hair pins in worked bone, a biconical terracotta spindle whorl and loom weight, a bronze city coin of Sagalassos depicting the lunar god Men and minted during the reign of Maximinus (235-238 CE), a fragment of a millefiori bowl, pieces of a miniature alabaster bowl, a miniature lead jug and fishplate, as well as a locally produced Early Roman Imperial miniature ceramic oil lamp and again some seashells. Furthermore, several fragments of human bone and some teeth were retrieved from these disturbed deposits. A very hard, 0.19 m thick deposit of greyish brown silty soil was found underneath these layers of excavated soil (Fig. 7). Its compactness and lack of intrusive (modern) material identified it as an in situ deposit. The deposit contained large amounts of finds which, due to their specialised nature, could be identified as an assemblage of votive offerings: it included several miniature lead vessels, fragments of lead miniature mirrors, pottery for dining, glass vessels such as unguentaria and bowls, and some tools, including a glass and worked bone distaff, and a miniature iron chopping knife. The majority of the finds from this deposit equally consisted of several hundreds of fragments of terracotta figurines: many different types of Aphrodite, Eros, Tyche, Athena, and Hermes were again predominant among the representations of deities, while female busts were prevalent among the human representations. Less numerous categories of votive offerings were terracotta plaques, such as one of Eros embracing Psyche, and lead figurines, like the fragmentary triad of the gods of medicine, Hygieia, Telesphoros and Asklepios. A bronze city coin of Perge minted during the reign of Hadrian (117-138 CE) and depicting the local goddess Artemis Pergaia was also found.

Fig. 7. In situ deposit (locus 15) in the SW part of Room 2 with numerous fragments of votive offerings (© Sagalassos Archaeological Research Project). 258 P. TALLOEN ET AL.

This ‘assemblage’ of votive goods could be generally dated to the 1st-2nd century CE based on the associated pottery. Yet, several elements indicated that it had not been deposited there originally: 1) the general fragmentary nature of the material, with none of the pottery or figurines, for example, being complete or broken in situ; 2) the weathered nature of the material which suggests that it had been exposed to the elements for a considerable period of time prior to being buried; 3) the presence of several fragments of terracotta roof tiles and water pipes, which clearly must have belonged to structures outside the crevice. Consequently, it was most probably the result of a clear-up operation within the sanctuary. Judging from the slope of the deposits, this material appears to have been thrown in from the north where the ‘entrance’ to Room 2 is situated. Few shards of 4th century CE pottery present within the deposit provide a terminus post quem for this operation. Underneath the layer with redeposited votive material, a hard, 0.05 m thick deposit of brown silty soil was revealed. This must have served as the substrate of a walking level, and was probably arranged in the 1st century CE, based on the post quem date provided by the ceramics retrieved from the deposit which also included some prehistoric shards. It had been laid out on top of a fill consisting of small and medium-sized limestone fragments which had been deposited between the parts of the bedrock in order to create a level area. The latter layer could not be completely excavated because of the bedrock and large limestone blocks between which it was present. It hardly contained any artefacts, but the presence of two almost complete human bones – a right shinbone (tibia) and calf bone (fibula) which showed a similar size and, although not found in anatomical position, may belong to the same adult individual – and a large base fragment of a prehistoric pot should be noted. This was the lowest stratum that could be excavated in this part of Room 2. In the NE Area of Room 2, the top layers with a thickness of 0.18 m, as well as the 0.85 m thick fill of a round pit (diameter: ca. 1.20 m) dug by looters in the central part of the area consisted of previously excavated soil. It included large amounts of pottery shards and fragments of terracotta figurines. The chronological range and typology of these artefacts was comparable to the material found outside the space, again confirming Room 2 as the origin of the debris. Underneath, a sequence of deposits, apparently untouched by the illegal excavations, was found along the edges of the space. The ca. 0.30 m thick top deposit, mainly contained fragmentary Roman Imperial period artefacts, including pottery and terracotta figurines, as well as fragments of glass vessels, hair pins and a spindle in worked bone. It was covering a more compact deposit with a thickness varying between 0.30 and 0.40 m of which the ceramic contents could be dated to the (late) 1st-2nd century CE with some residual Late Hellenistic material. The more exceptional finds included an almost complete terracotta figurine of Aphrodite anadyomene and of Tyche standing within her shrine, a finger cymbal, a fragment of a worked bone spindle and a complete oil lamp with traces of soot on its nozzle. Yet, the presence of roof tile fragments as well as few 4th century CE shards, as encountered in the SW Area, again point to remains of a Late Roman clean-up operation rather than a votive deposit. The dump was preceded by a 0.14-0.24 m thick deposit of silty soil with many charcoal remains, as well as fragments of terracotta figurines and pottery dating to the 1st century BCE and 1st century CE. The head of a terracotta figurine representing a comedy actor, A PLACE ON THE FRINGE OF SAGALASSOS 259 datable to the 2nd-1st century BCE, is one of only few fragments of Hellenistic figurines found at the site. Within the deposit were concentrations of shards belonging to several Early Roman Imperial common ware vessels, including a cooking pot and a jug, though none constituted a complete vessel. It covered a level of deposition on top of which shards of two complete, locally produced unguentaria of the spindle bottle type, dating to the 2nd-1st century BCE, were retrieved, as well as an iron cauldron fork; parallels for this type of fork have been found in Gaul where they were dated to the late 2nd-1st century BCE9. Given the completeness of the unguentaria – the only items found within the complex to be broken in situ – they were probably in their original location of deposition. The substrate of this level of deposition was a 0.25 m thick layer of silty soil with charcoal remains and many small stones which included three human cervical vertebrae, likely from one adult individual. The lowest investigated level, a deposit of greyish silt that had accumulated between large and medium-sized stones, could not be completely excavated due to the presence of the large stones. It contained no finds and may therefore have been a natural accumulation.

Middle Zone The Middle Zone located in the centre of the outcrop consisted of two spaces, Rooms 3 and 4, that could only be partially investigated. Room 3, situated on the northern edge of the outcrop between Rooms 1 and 2, is a roughly rectangular space of which only the eastern part, measuring 1.06 m (N-S) by 1.86 m (E-W) could be investigated due to the collapsed ceiling. It too had a natural opening between the covering limestone boulders, comparable to that of Room 2, that may have served as an access to the space. Based on the sloping stratigraphy of the debris of illegal excavations present on top of the ancient walking level in Room 1 (see above), Room 3 could be established as the original location of the excavated material. This may explain why only few deposits were found in this part of the complex. The top layer was a 0.15 m thick deposit of previously excavated material. Other than pottery, it again included numerous fragments of terracotta figurines. The represented types included Aphrodite anadyomene, Aphrodite strophion, Aphrodite of Knidos, Aphrodite seated on a high-backed chair, Athena, Kybele seated on a throne flanked by lions, Tyche holding a rudder and cornucopia standing within a columnar shrine, Nemesis, Eros holding Psyche, Hermes, a kourotrophos or nursing woman, and busts of women. The underlying layer of compact silty soil with a thickness of 0.11 m appeared to be in situ and contained pottery and fragments of terracotta figurines that could be attributed to the 1st century BCE-1st century CE. It was deposited on top of the surface of a compact deposit of silty soil (locus 8) with many small limestone fragments and a thickness varying between 0.12 m and 0.22 m. The few ceramic finds, only present in the top part of the layer, could be attributed to the Early Roman Imperial period, providing a terminus post quem for the arrangement of this level of deposition. Underneath, the limestone bedrock was present.

9 Perrin 1991. 260 P. TALLOEN ET AL.

Room 4, a natural cavity with trapezoidal plan, 2.40 m long (NW-SE) and 1.40 m to 2.30 m wide (SW-NE), is equally situated in the central part of the complex, between Room 1 (to the east) and Room 2 (to the west), and at a level below Room 3 (to the north). This room was originally located at the western end of the corridor (Room 1) but the collapse of the ceiling eventually separated both spaces (Fig. 8). Unfortunately, the illicit excavations that occurred during the last decades of the 20th and the beginning of the 21st century, also took place in this part of the complex. They have rendered the stratigraphy of the space largely illegible because it now lies buried beneath thick deposits of large limestone blocks that were piled up there by the looters and cannot be removed through the small openings leading to the space. A 1.15 m thick deposit of previously excavated material was found throughout the space. In addition to numerous fragments of pottery and terracotta figurines dating to the Roman Imperial period (mainly 2nd century CE), as well as some pieces of building ceramics, it included an almost complete terracotta figurine of Eros and Psyche embracing each other, and an iron sewing needle. As could be told from the orientation of its sloping surface, the excavated material was deposited into Room 4 from Room 3, the higher situated space to the northeast of Room 4. As this was also the space from which the material found on the surface of Room 1 originated, it too appears to have been a dump for the clean-up of the sanctuary, similar to Room 2. Only in the northeastern corner of Room 4 could in situ deposits be reached. Firstly, a 0.17 m thick fill of silty soil was encountered which included early 1st century CE ceramics and some butchered bones of cattle. It was on top of this deposit that part of the limestone ceiling fell, closing off Room 4 from Room 1 and preventing access to the space from the corridor. Underneath, a floor substrate of hard clayish soil with many charcoal remains was found, arranged sometime during or after the 1st century CE according to the ceramics it contained. The substrate could only be partly excavated due to the presence of large limestone blocks. This was the lowest excavated deposit in this space.

Fig. 8. View of Room 4 Fig. 9. Late Chalcolithic pottery from RS and its collapsed ceiling from the west (© Sagalassos Archaeological Research Project). (© Sagalassos Archaeological Research Project). A PLACE ON THE FRINGE OF SAGALASSOS 261

THE OCCUPATIONAL HISTORY OF THE SITE

During its use, RS was a ‘place’, as a culturally significant locale that existed within a landscape and was meaningful to specific cultural groups through everyday experience and shared stories associated with them10. Throughout history, human beings found cave-like spaces alluring, often as a place of divine immanence and a zone of contact with the under- world, but also fulfilling several other functions11. In order to understand the significance of this cave-like site in the ancient landscape of Sagalassos, we need to look not only at its location and form, but also at the artefacts that were deposited there. While the general appearance of the site remained largely unchanged throughout the ages, it is the artefacts that can inform on the nature and chronology of its usage. Artefacts are of crucial importance for understanding the function of the locale. The presence of specific artefact categories, the relevance between various artefacts, their specific contexts, and the manner of their deposition, all manifest social action and provide the particular character of the site12. Analysis of the material culture from the excavated contexts situated the main use of RS between the early 2nd century BCE and the early 3rd century CE; few 4th century CE shards present among them indicated that people still visited the site in late antiquity. The presence of some Middle Byzantine ceramics, most of which were associated with a probable hearth, pointed at the renewed occupation of the crevices, after a gap of many centuries. Interestingly, residual prehistoric shards dating to the Late Chalcolithic period were also found among the excavated assemblages. These three different chronological phases of usage of the site will now be discussed in further detail in order to understand the nature of its occupation during those different periods.

A Chalcolithic burial site? In addition to the many categories of votive offerings that were found at the site, perhaps the two most surprising elements were the discovery of human remains and Late Chalcolithic pottery. Some of the 38 shards of hand-shaped pottery could be fitted suggesting that they belonged to a limited number of vessels that were deposited at the site, including a cup and a bowl (Fig. 9). These shapes have been dated elsewhere to the Late Chalcolithic period (4000-3200 BCE)13. Other than the aforementioned two almost complete bones, the retrieved human bone material consisted of 222 unburnt bone fragments, 25 teeth and 28 burnt bone fragments. The majority could not be assigned to a specific skeletal element. Most identified fragments were derived from the lower limbs, with a minority from the upper limbs, the skull, spine and the pectoral and pelvic girdle. One tooth could be identified as non-adult. Most bones were small long bone diaphyseal fragments. Age could not be estimated with certainty, although the size of the bone fragments suggested they did not belong to children; also sex could not be determined. The right

10 Harmanşah 2014: 1. 11 Mavridis et al. 2013. 12 Mavridis et al. 2013: 2-3. 13 Duru 2008: 135-141. 262 P. TALLOEN ET AL. proximal femur, the right distal fibula and the right second metacarpal showed a minimum number of individuals (MNI) of two, while the right distal anterior tibia showed a possible MNI of three. The presence of a non-adult molar also suggests an MNI of three, with one non-adult in addition to two adults. Unfortunately, nearly all these bones and prehistoric shards were found out of context, in previously excavated deposits both within and outside of the natural cavities. Yet, the presence of two almost complete bones of a lower leg in the lowest excavated deposit in the SW Area of Room 2, found between large limestone blocks together with a large fragment of a Chalcolithic bowl, suggests that this was probably the original place of deposition which had been disturbed by later usage of the space. The levelling deposits with Chalcolithic ceramics found in Room 1 can perhaps also be linked to the prehistoric use of the cave-like complex. They were most probably arranged there to facilitate the access to the inner spaces, although it cannot be ruled out that this occurred at a later point in time with the shards present in those deposits as residual material. Nevertheless, a prehistoric occupation phase of the site is certain. The large amount of human bone found in deposits inside Room 2, as well as in the dumps of the illegal excavations outside of that space suggest that it originally served as a place for entombment. The use of the cave-like spaces as part of a sanctuary during the Hellenistic and Roman Imperial periods rules out that the human remains can be attributed to this phase of usage, as interment within sanctuaries was strictly forbidden during classical antiquity14. Furthermore, the presence of in situ late antique deposits on top of some of the bones and pottery rules out a later date as well, which brings us to the earlier use of the site during the Chalcolithic period. The variety of ways to bury the dead expanded significantly during this period throughout the Mediterranean, with caves and subterranean spaces becoming common places for the dead15. Caves as places of burial during the Chalcolithic period are also attested in southwestern Anatolia as illustrated by the burials in the cave of Öküzini near Antalya16. The combination of skeletal material together with Chalcolithic pottery in one of the lowest excavated deposits is therefore interpreted here as an indication of the use of the crevices as a burial ground during the 4th millennium BCE. Given the small number of vessels and individuals, this usage appears to have been brief. In any case, the assemblage represents the earliest traces of human settlement in the immediate vicinity of the later city of Sagalassos and can undoubtedly be linked to the increased settlement activity in the wider area of the Ağlasun Valley during the Late Chal- colithic period17.

14 Mikalson 2004: 8. 15 Rowan 2018: 134-137. 16 Kartal and Erek 1998. 17 Vandam et al. 2017. A PLACE ON THE FRINGE OF SAGALASSOS 263

A Hellenistic and Roman Imperial period sanctuary The lack of subsequent pre- and/or protohistoric deposits indicate an interruption in the occupation of the site until the Hellenistic period. The high quantity of Hellenistic material dating to the 2nd and 1st centuries BCE, albeit found mainly in later, Roman Imperial period deposits, indicates that the cave was again used from that time onwards. The great majority of the pottery retrieved at the site is of local production and can be safely dated between the 2nd century BCE and the early 3rd century CE, with a clear focus in the 1st centuries BCE and CE18. Yet, the fact that much of the stratigraphy was disturbed by illicit excavations does not allow a more precise date for most loci, and even in situ deposits display this relatively wide chronological range of pottery. Some fragments clearly predate the 1st century BCE: several vessel shapes were identified that are strongly tied with earlier Hellenistic if not pre-Hellenistic traditions. At the other end of the main period of use of the sanctuary, there is a considerable amount of 2nd century CE pottery. Moreover, there is a small quantity of the locally manufactured Sagalassos Red Slip Ware (hereafter SRSW) that typologically belongs to the 4th century CE, material that can most probably be related to the end of the sanctuary (see below). As suggested by the different categories of specialised material culture that could be identified as votive offerings, it now assumed another function, becoming a cult site. Interestingly, this establishment of the sanctuary occurred shortly after the community of Sagalassos developed into a city-state and is perhaps related to the place of different social groups within this new constellation19. Since elements of material culture are held to reflect cultural values and religious beliefs that are understood to prefigure them, the materiality of cult will be used here as a basis for reconstructing rituals, establishing the concerns that they addressed, and ultimately identifying the people involved. Yet before turning to that, first the ritual space of the sanctuary will be addressed.

Ritual space In the landscapes of the ancient Mediterranean, natural formations like caves and other types of cavities were often thought of as places beyond the world of the living. As naturally numinous places that embodied the powers inherent in nature, people were drawn to them and established sanctuaries there, giving the natural landform a cultural significance20. As mentioned above, the distinctive topographical feature at the heart of RS is a cave-like space, consisting of large crevices in the rock outcrop which are covered by huge blocks of limestone, somewhat similar – mutatis mutandis – to the situation of the better-known sanctuary of Kapıkaya in the territory of Pergamon21. Except for the possible, minor interventions noted near the entrance in the southeast part of the outcrop, neither the outcrop nor any of the investigated spaces displayed any

18 Poblome et al. 2018. 19 Talloen and Poblome 2016: 117-120; Talloen 2019a: 191-193. 20 Bradley 2000: 25-27; Mavridis et al. 2013: 1. 21 Nohlen and Radt 1978. 264 P. TALLOEN ET AL. signs of an attempt to modify and shape them architecturally. Outside of the cavities, there are no signs of any significant architectural modification, although the substantial amount of roof and cover tile fragments recovered from the site does suggest that one or more places in the sacred area of the sanctuary were roofed. These were perhaps small ‘picnic’ spots where the ritual dining took place (see below). Since all sacrifices generally had to be consumed within the sanctuary22, the presence of appropriate cooking facilities and a suitable eating place is to be expected. Yet, the absence of other durable building material, such as carved stone blocks or bricks, suggests that they were only semi-permanent structures in mainly perishable materials. Also the fragments of large storage vessels or pithoi point to installations being present in or around the sanctuary, as these vessels were unlikely brought by people visiting the sanctuary. Possibly they contained water for use during actions in the sanctuary, or otherwise could have supplied visitors with drinking water. Several pieces of terracotta water pipes equally suggest the presence of a water installation of some kind. As purification by water upon entering a sanctuary was a standard ritual necessity23, the existence of such an amenity would not be surprising. Within the crevices only the laying out of floor levels (or levels of deposition) through the arrangement of fill deposits could be established. Especially the Roman Imperial period, more specifically the 1st and the early 2nd centuries CE, saw the addition of several floor levels in Rooms 1 and 4, which suggests an intensive use of the cult site at this time. Unlike the Roman deposits in Rooms 2 and 3, the in situ deposits in Room 1 could be characterised as floor levels which did not contain large amounts of votive material, corroborating the identification of the space as a corridor or passage way to the actual centre of the cult site. The latter was most probably situated in the centrally located Room 4, where (some of) the rituals could have taken place. Altar(s) and cult image(s) may originally have been present in this central space of the outcrop, although a natural feature could have been used instead. Whatever its exact shape, it is clear from the presence of several used oil lamps found among the votive offerings (Fig. 10) that this central space will have been dark and cave-like, and therefore numinous in the mind of the ancient population, a place where humans could communicate with the divine world through religious ritual. So, experience of the inner part of the cavern will have been an essential part of the rituals that took place there. Overall then, the appearance of the site is that of an unaltered feature of the natural landscape. There are no traces of monumentalisation, nor any representational (e.g. statu- ary or rock-cut reliefs) or epigraphic content (i.e. public or private inscriptions). Not even any obvious modifications to meet the requirements of worship, like rock-cut steps, benches or votive niches were recognised. The natural place was turned into a cultural place simply through the deposition of artefacts (see below). All this identifies it as a ‘natural sanctuary’, not a monument constructed by human labour but a non-monumentalised cult site with a natural feature – in casu a cave-like crevice – as primary recipient of worship24.

22 Dignas 2007: 173. 23 Bendlin 2007: 181. 24 Bradley 2000: 34; Mylonopoulos 2008. A PLACE ON THE FRINGE OF SAGALASSOS 265

Fig. 10. Roman Imperial period oil lamps from RS (© Sagalassos Archaeological Research Project).

Although the civic community was undoubtedly knowledgeable about its natural envi- ronment and thus aware of the existence of the rock outcrop and its cavities, something that is corroborated by the fact that the sanctuary already came into existence shortly after Sagalassos became a city, at no point during its more than 400 year long life was there any attempt to monumentalise the cult site. Even when the site experienced a boom in popularity during the Early Roman Imperial period, as evidenced by the number and quality of the votive offerings left there (see below), this did not translate into a modification of the natural setting. All this indicates that there was a deliberate choice not to invest in it architecturally. Consequently, it did not come to dominate the (sacred) landscape and assume a more powerful role in the life of the civic community, unlike other natural sanctuaries in the region such as Arpalık Tepe near Selge, Eleksi Tepe near Mallos, İnarası near Keraia, and Zindan Mağarası near Tymbriada25. Such a monumentalisation of natural cult sites can be seen as part of the political appropriation of particular local practices by the ruling elite, which introduced monumentality and public spectacles to these already significant sites of cultural practice while they were used by different cultural groups26. Con- versely, the absence of signs of official involvement – not only monumental architecture, but also civic priesthood and monumental writing – would situate the sanctuary outside of the sphere of public/polis religion27. RS can therefore be characterised as a (private?) cult site that was established by ordinary people and remained a focus of popular cult throughout its existence.

Ritual acts In spite of the lack of monumental markers, RS retained its sacred function over several centuries and knowledge of the place was handed over from generation to generation.

25 For an overview see Talloen 2015: 236-240; for İnarası near Keraia see Ekinci and Zenger 2017. 26 Harmanşah 2014: 3. 27 Kindt 2009. 266 P. TALLOEN ET AL.

Although commemoration generally occurs through the construction of monuments, another way of creating memories is through ritual acts28. For such an event to be remem- bered, it had to be an active process conducted between successive generations of people if it were to have any importance, as there were no obvious visible reminders29. For the case of RS, this continuance is brought out by the uninterrupted deposition of objects between the early 2nd century BCE and the early 3rd century CE. Such persistence hints at a shared body of knowledge that was part of local identities. Although most of the examined contexts had fallen victim to illicit excavation, the observed consistency of the material assemblages throughout disturbed and undisturbed contexts was remarkable. This demonstrates that the deposits from RS are still representative for the original composition of the votive assemblages and therefore valuable for archaeological research. Judging by the material assemblages retrieved from the cult site, two kinds of ritual behaviour occurred throughout its history: the consumption of ritual meals and the deposition of votive offerings.

1) Ritual meals

Both the dumps of the illicit excavations as the in situ deposits contained finds – namely pottery and faunal remains – that could be identified as the debris of meals which were held at the sanctuary throughout its lifespan. What pottery did people bring with them when they visited the cult site? In terms of function, since the majority of fragments concern closed and open forms of locally produced tableware30, we can envisage a place where they came to dwell on and celebrate specific occasions, and to that purpose brought food and beverages to do so. Two main functional groups of pottery can be distinguished. The first consists of slipped table wares, vessels used for the consumption and serving of food and beverages: plates, bowls, dishes, cups, but also smaller and larger jugs. The earliest pottery related to such meals has been dated to the first half of the 2nd century BCE, and local versions of vessels with West Slope-style painted and incised decoration (Fig. 11a) also likely belong to the second century BCE31. Typical shapes that belong to the Late Hellenistic phase more generally include plates with upturned rims (Fig. 11b), mastoi and related drinking cup shapes32, various bowls and dishes, lekanè or lekanè-like vessels (for preparing wine?)33 as well as the occasional down-turned rim plate. Fishplates and mouldmade (‘Megarian’) bowls were not identified. The Early Roman Imperial evidence is to an important extent represented by a range of SRSW types34. Cup types 1A100, 1A110-1, 1A130, 1A160-1 and to a lesser extent 1A150 are very well represented and, given their date ranges and typological appearance, represent successive stages of use of RS during the first two

28 Bradley 2000: 157. 29 Bradley 2000: 158-159. 30 More generally also see Van der Enden et al. 2014. 31 Rotroff 1991. 32 Van der Enden et al. 2018. 33 Daems et al. 2019: 86-88, fig. 2. 34 For detailed discussions of the various types of SRSW mentioned, see Poblome 1999. A PLACE ON THE FRINGE OF SAGALASSOS 267

centuries CE. This timeframe is further corroborated by bowls (mostly types 1B170 and especially 1B190-1) and dishes (types 1C100 and 1C120-3 in particular). A rather broad range of other SRSW bowl, dish and plate types are present but usually in smaller numbers. Lekanè type 1F150 was, on the other hand, fairly common. The second group concern vessels in fabrics that are suited for storage and cooking, although the question remains whether people visiting the sanctuary frequently prepared food on the site. The fair quantity of closed vessels, as well as the relatively limited number of actual cooking vessels, suggest that people presumably brought their meals and beverages to this place rather than preparing them there. Nevertheless, this should not completely be excluded given the presence of cooking utensils among the finds, such as the Late Hel- lenistic cauldron fork used for the boiling of meat in a cauldron, a part of the obligatory

b c

Fig. 11. Hellenistic ceramics from RS: a. two fragments with incised-and-dot painted West Slope- style decoration, datable to the Late Hellenistic phase of RS. The fragment on the right belongs to a lekanè; b. fragments (giving a complete profile) of an upturned-rim plate with a partly preserved stamp at its centre, presumably datable to the (middle/second half of the?) 1st century BC; c. a shallow and small-sized dish (miniature?) (© Sagalassos Archaeological Research Project). 268 P. TALLOEN ET AL.

Fig. 12. Partly restored bases of unidentified (closed) vessels of local manufacture, possibly kraters (© Sagalassos Archaeological Research Project). equipment of the sacred cuisine35. The relatively low number of faunal remains related to consumption, which stands in contrast with the large amounts of pottery related to dining, also suggests the preparation of meals outside of the sacred precinct. Together with the lack of charred remains among the animal bones representing altar-burnt sacrificial materials, this also indicates that animal sacrifice will not have been a very common or important ritual at RS. The meat of such sacrifices generally had to be consumed within the sanctuary, which in turn would have yielded a far greater amount of animal bones than actually retrieved. The excavated faunal remains related to consumption mainly consisted of chicken, the cheapest (sacrificial) animal but also the one favoured by certain gods, such as Aphrodite and Asklepios36; skeletal elements of sheep and goat – especially canon bones (metapodials) – were also represented, as well as some pig and cattle. Among the common ware vessels are about a dozen of reconstructed bases (Fig. 12), known also from contexts within the urban area of Sagalassos where they date to the Early Roman Imperial period. These can be classified under one type of vessel, possibly krater- like used for mixing wine with water, and thus hinting at the consumption of wine at the cult site. The overall repertoire of the pottery as well as its quality, in comparison to other contemporary excavated contexts from the urban centre, basically represents nothing out of the ordinary. The ceramic picture that emerges from these two categories is that which one might expect in a household. In other words, what people brought along presumably came from their own households. This could explain the observed limited functional variety on the one hand, and the rather broad morphological range on the other. What was brought to the sanctuary had an internal logic and was likely related to the number of people in one specific group, as well as to what would be consumed. These various functional categories allow us to picture how an assemblage used during a ritual meal in the sanctuary – theoretically – could have looked like (Fig. 13), which very

35 Durand and Schnapp 1989: 56. 36 Villing 2017. A PLACE ON THE FRINGE OF SAGALASSOS 269

Fig. 13. Partly restored locally manufactured vessels, a theoretical reconstruction of the main functions/vessel types used in a ritual meal during the Early Roman Imperial period (© Sagalassos Archaeological Research Project). likely included bowl type 1B191 (bottom left) and cup types 1A130 (bottom right) and 1A161 (centre right). The partially reconstructed vessels, a chytra (top left) and some form of lopas – a low, shallow vessel used for frying or stewing (centre bottom) could have been used for preparing food, though, as pointed out already, previously prepared food was likely brought, quite possibly in such and/or other vessels. However, the lopas-like vessel, possibly a cauldron, is of sizeable dimensions and filled with food may have been unprac- tical to carry. It is a rare find of some interest: combined with the iron fork mentioned above provide tentative albeit tantalising clues that people may have enjoyed a fondue avant-la-lettre. No large sets of dining accoutrements, like those recently discovered at a cave sanctuary in Pergamon37, have been retrieved at RS. Together with the lack of other indications for some form of communal organisation such as the arrangement of sizeable dining spaces, this suggests that the meals held at the cult site appear to have involved smaller groups of people, bringing their own crockery, not the celebration of communal festivals. Such ritual meals were typical occasions at which the common identity of the group was reaffirmed38. So, we could possibly envisage family groups who came to the sanctuary to ritually celebrate an important social event in their private lives and enjoyed a meal together in small makeshift shelters outside of the cavities. In contrast with the numerous local products only a tiny quantity of imported pottery could be recognised, which thus far concerns six fragments (i.e. 0.04% of all pottery) and likely an equal number of individuals (Fig. 14). This includes three lead-glazed vessels, including one skyphos (top right) datable to the (mid-)1st centuries BCE-CE. Although Perge in Pamphylia is a likely source, the (macroscopic) fabrics of these three fragments show some variation. Another fragment originates from southern Lycia, quite possibly from the vicinity of Limyra (where it is extremely common), a category of cooking and

37 Engels 2015; 2019. 38 Mylonopoulos 2006: 83-84. 270 P. TALLOEN ET AL.

Fig. 14. Fragments of non-regional provenance: Top row, Lead-Glazed vessels (a skyphos base top right) one or more presumably manufactured in Perge; Bottom right, a handle fragment of a vessel manufactured in the area of Limyra (southern Lycia) (© Sagalassos Archaeological Research Project).

utilitarian wares (‘pâtes lyciennes kaolinitiques’)39 that is recognised now and then in Saga- lassos. The meaning of these handful of imports is not entirely clear. It is difficult to imagine that such imported vessels (and their rarity) – certainly the usually green-and- yellow lead-glazed examples – escaped the attention of the inhabitants of Sagalassos, and an ‘exotic’ connotation may have enhanced their cultural significance. On the other hand, they also feature sporadically among assemblages found in the urban centre and may therefore simply reflect what was available at the time. Upon viewing the pottery evidence from RS as a whole, in particular the SRSW, one can hardly escape noting its somewhat motley appearance. This heterogeneity concerns slip colour (a range of colours and a mottled appearance being more common features for the Late Hellenistic period), for example, and also the quality – a judgment that inevitably rests on thin ice – does not generally represent the top end of what the local potters were capable of during the Roman Imperial period. There is of course our broad chronological perspective, and the pottery has obviously suffered from various human and natural phe- nomena during and after the period of use of RS. Yet, this perceived heterogeneity can tentatively be brought in relation to the status and purpose of the sanctuary or could also hint at a certain pragmatism: knowing they had to leave behind their utensils upon leaving the sanctuary people may have chosen not to bring their finest.

2) Votive deposition In terms of material residue, a religious practice that was even more important than the consumption of meals was the ritual of votive offering, especially from the Early Roman Imperial period onwards. Fundamental to the study of this ritual is the identification of material objects that were brought to the site as votive offerings. Because of the many situations in which they were used and the various functions they fulfilled, but also depending

39 Lemaître et al. 2013. A PLACE ON THE FRINGE OF SAGALASSOS 271 on the socio-economic status of the votaries, votive objects could take on many forms40. As a result of standard shapes and imagery, the cultic nature of certain types of evidence like representations of deities, is fairly easy to establish. In many cases, however, ritual objects cannot be recognised as easily. Such objects were endowed by ritual with qualities that brought them into relation with the divine or made them conducive to the efficacy of the ritual, and thus caused them to enter the domain we label ‘sacred’, while there is nothing in their intrinsic nature that distinguishes them from objects of everyday use, like the implements used in cooking or pieces of personal adornment. Except when present in an explicitly cultic context such objects often remain undetectable. For example, the over- whelming number of terracotta figurines – literally thousands of fragments – that were found at RS, constitute a clear specialisation in the deposition of objects with a known ritual function that allowed it to be identified as a special-purpose site41. Consequently, other material categories from the same stratigraphical contexts can be interpreted in the same ‘ritual light’, as the figurines were contextually connected with this material culture, entangled within a web of materiality. Yet, it remains methodologically challenging to distinguish some votives from the artefacts that were used for ritual meals. A full overview of the finds from RS will be presented elsewhere; it will suffice here to list the categories of offerings and only go into further detail for the most important ones. During the Hellenistic period, ritual remains other than pottery used in the ritual meals at the site were restricted to some glass and ceramic unguentaria or ointment bottles, as well as a handful of fragments of terracotta figurines. The celebration of the event with meals was obviously more important at the time than the offering of gifts, unless the latter were of perishable nature and left no traces in the archaeological record. The deposition of votive offerings increased dramatically during the Roman Imperial period. These now included purpose-made offerings such as lead and terracotta figurines, terracotta and metal plaques, miniature vessels in stone, glass, lead and terracotta, as well as miniature tools and mirrors. Furthermore, objects of the instrumentarium domesticum such as worked bone hairpins, pestles, spindles, whorls and distaffs, glass vessels (including a millefiori bowl), glass and ceramic unguentaria, stone cosmetic palettes for mixing makeup, metal and glass pieces of jewellery, ceramic oil lamps, metal finger cymbals, styli, and a few coins were found. Finally, the group of Mediterranean sea shells should also be counted among the votive gifts, since these were commonly dedicated to goddesses, especially to Aphrodite who had a close connection with the marine world42. A first significant category of votive gifts, in terms of quantity, comprises miniature vessels, a part of the pottery repertoire that has thus far not been encountered anywhere else in the urban area of Sagalassos and is therefore unique to RS, which further underlines the time- and place-specific character of the site (Fig. 15). Like other types of miniatures unearthed at RS, they can be distinguished as votive gifts on the criterion of size43. Some of these represent miniature versions of SRSW types that are well-attested, both in RS and

40 Osborne 2004. 41 Osborne 2004; Alcock and Rempel 2006. 42 Theodoropoulou 2013: 210-211. 43 Barfoed 2018. 272 P. TALLOEN ET AL.

Fig. 15. Miniature vessels in the locally manufactured Sagalassos Red Slip Ware, with fragments of cup types 1A160-1 on the right, and a lobed dish middle left (© Sagalassos Archaeological Research Project). urban Sagalassos, such as cup type 1A160-1 (pictured in the right half of Fig. 15; Fig. 16a) the most common miniature type attested. Also identified are miniature versions of cup type 1A150 (Fig. 15, top left), bowl types 1B162 and 1B190, and dish type 1C130-3; the latter types are represented by only one or two specimens. It appears that whilst some miniatures were wheel-made, others were obviously handmade. The characteristic ledge handle of SRSW that normally was mould-made and subsequently applied horizontally onto the rim was on occasion also miniaturised and ‘crudely’ incised (unrecognisable here) in imitation of their mould-made counterparts (Fig. 16b). Although the date ranges of their ‘originals’ extends into the 3rd century CE, we presume these miniatures belong with the main phase of Roman Imperial use of RS, the 1st and 2nd centuries CE. Of particular interest is a partly restored miniature version of a lobed dish with incised rim (Fig. 15, centre left) – a ‘normal-sized’ version in SRSW is known from a 2nd century CE context from the Urban Mansion. This shape is extremely rare at Sagalassos and was possibly manufactured per individual order. A last group are miniature vessels made in a fabric (Fabric 237) which was previously attributed to Düzen Tepe, a site ca. 1.8 km southwest of Sagalassos that was occupied since Late Achaemenid times and eventually abandoned in the earlier 2nd century BCE. They appear as small carinated dishes with string-cut disc bases (Fig. 11c) that are not unlike more squatted, shallow versions of Achaemenid bowls44. Even if these miniature versions of pottery can hardly have served a practical purpose as part of a ritual meal, they may nonetheless have contained a bit of food or liquid when they were deposited as part of a ritual. The purpose of miniature vessels has received quite a bit of attention in recent years, including attempts at a classificatory system45. One of the results was that their purpose and meaning extended well beyond the notion that these were simply smaller and therefore more practical objects to carry and deposit46. Indeed, the quantity of miniature vessels at RS pales in light of the quantity of figurine fragments

44 Daems et al. 2019: 85-86, fig. 2. 45 Barfoed 2018. 46 Martin and Langin-Hooper 2018. A PLACE ON THE FRINGE OF SAGALASSOS 273

a b

Fig. 16. a. A handmade miniature version of cup type 1A161, decorated with stylised bunches of grapes?; b. A miniature ledge handle with incised decoration (© Sagalassos Archaeological Research Project).

(see below). Given that RS partly functioned in relation to the upbringing and well-being of children (see below) these miniatures were perhaps a reflection of that element. This is not to say that these were originally toys – though they could have been – but that their size was a material reflection of the ‘child dimension’ of RS. Miniatures may in fact have appealed to the sense of perspective and imaginative world of children – as it (still) does today – more so than the normal-sized ceramic repertoire. Locally produced, mould-made terracotta figurines constitute by far the most numerous category of votive offerings at RS during Roman times. Most of these terracottas survive in small fragments – more than 40,000 yielded by the rescue excavations alone – and only a limited number is better preserved, while none survives complete. The fragmentation of the figurines, like that of the pottery and most other finds from the sanctuary, is undoubtedly due to a number of factors but, above all, the clean-up during antiquity (see below) and the later robbing of the site must have been responsible for their poor state of preservation. A conservative estimate, based on the patterns of fragmentation, suggests a total of more than 3000 figurines, the largest number to be documented so far in a single site anywhere in Anatolia. As cheap but specialised objects, affordable to all classes of society, terracotta figurines were a popular category of votive gifts in most regions of the ancient Mediterranean47.

47 Huysecom-Haxhi and Muller 2015. 274 P. TALLOEN ET AL.

Given that only few fragments of (Late) Hellenistic figurines could be registered, the finds from RS point to the predominant use of such images during the 1st and 2nd centuries CE, a development that can most probably be related to the start of their local mass production during the Early Roman Imperial period48. In the contexts that have been studied so far, ca. 40 % of the identifiable terracotta figurines consisted of representations of human individuals, belonging to different age classes (Fig. 17): young girls (depicted as students holding tabellae or writing tablets and as busts; Fig. 17d), young women playing musical instruments such as the harp and lyre (Fig. 17c), nurturing mothers or kourotrophoi (Fig. 17b), and especially shoulder busts of adult women (up to 80 % of all human images; Fig. 17a). Male figures, on the other hand, are only attested as boys (again as students and as busts; Fig. 17e) and adolescents (on horseback or as athletes; Fig. 17f), not as adults. Although these images are generic, not realistic portraits, they should probably be seen as representations of the votaries or the individuals for whom divine protection was invoked49; in the case of RS: women, children and youths. Representations of deities were most common, making up some 59 % of all terracotta figurines. Among them figurines of Aphrodite proved predominant throughout all studied find contexts (69 % of all figurines of deities; Fig. 18a-d), followed by her son Eros (10.5 %) who was frequently accompanied by his girlfriend Psyche (Fig. 18e-f). Other female deities included (in order of frequency) Tyche, Athena, Nemesis, Kybele, Isis and Hygieia (Fig. 19a-c). Overall, male divine presence was restricted, with Hermes being the most represented god (4 %); others comprised Ares, Asklepios, Harpokrates, Herakles, the moon god Men, and Sarapis, all of whom were attested by only few examples (Fig. 19d- f). The absence of the leading civic gods of Sagalassos, Zeus and Apollo50, is striking. Figurines of animals are limited overall (less than 1 %). Most frequently attested are terracotta birds such a cocks and pigeons, the favourite (sacrificial) animals of Aphrodite51. In absence of inscribed votive offerings, the identity of the god(s) worshipped at RS can be approached through the identification and quantification of divine representations, as the proportion of the different typological categories of deities can be held to represent their relative importance at the cult site. Although it is not always the case that the tutelary deity of a sanctuary is represented by the largest number of figurines52, the outright dom- inance of her representations, together with the corroborating evidence of animal images (i.e. pigeons) and marine shells, clearly single out Aphrodite as the main subject of worship at RS. This is somewhat extraordinary as there is normally a significant relationship between the physical configuration of places where sanctuaries were established and the gods to whom they were dedicated53. Although commonly associated with lush gardens with fruit trees and flowers, cave sanctuaries for Aphrodite are scarce54. This suggests,

48 Talloen 2020. 49 Huysecom-Haxhi and Muller 2007. 50 Talloen and Waelkens 2004. 51 Villing 2017. 52 Alroth 1989: 112. 53 Bradley 2000: 25-26. 54 Bumke 2015. A PLACE ON THE FRINGE OF SAGALASSOS 275

a b

c d

Fig. 17. Terracotta figurines of human figures: a. female bust; b. kourotrophos; c. woman playing the harp; d. fragments of seated girls and boys with tabellae on their lap; e. fragments of busts of boys wearing a chlamys; f. fragment of a rider on horseback (© Sagalassos Archaeological Research Project). e f 276 P. TALLOEN ET AL.

a b

c d

Fig. 18. Terracotta figurines of Aphrodite and Eros: a. Anady- omene; b. Genetrix; c. Strophion (SA-1999-Y-6; height: 0.095 m); d. Pudica; e. two wrestling Erotes (SA-1999-Y-8; height: 0.07 m); f: Psyche and Eros seated on a throne, holding a pigeon (© Sagalassos Archaeological Research Project). e f A PLACE ON THE FRINGE OF SAGALASSOS 277

a b

c d

Fig. 19. Terracotta figurines of other deities: a. Athena; b. Tyche/Isis and Nemesis; c. Hygieia; d. Hermes; e. Ares; f. Harpokrates (© Sagalassos Archaeological Research Project). e f 278 P. TALLOEN ET AL. together with the fact that the goddess seems to have reached Pisidia only by the end of the Hellenistic period55, that she was probably a later addition to a sanctuary that may originally have served the cult of another god(dess). Cave-like locations are often predes- tined for the cult of the Mother Goddess who, because of her strong ties to nature, was predominantly worshipped at naturally formed cult places56. Yet, contrary to other cave sanctuaries in the region, such as Zindan Mağarası (Meter Theon Vegeinos), Kocain (Meter Oreia), Karain (Meter Oreia), Eleksi Tepe (Meter Malene?), Dutalan (Tyriose), and Arpalık Tepe (Meter Theon)57, clear evidence is lacking for RS. It is obvious from her limited number of terracotta figurines, however, that by the Roman Imperial period at least, Kybele only played a secondary role here and was overshadowed by Aphrodite. The broad repertoire of divine representations retrieved from RS would suggest that the sanctuary was not exclusively the preserve of a single deity but that several deities were worshipped there. Having said that, the observed multiplicity may also be (partly) explained by the phenomenon of »visiting gods«: images representing one deity which were given to another deity, because of a special relationship between the gods or as a pleasing gift58. This phenomenon could account, for example, for the singular presence of certain deities such as Ares that would then have been given as a pleasing gift to Aphrodite. Yet, it has also been understood as indicating that people dedicated what they had at hand and that it did not matter much what they gave59. Such a simplistic approach to the agency of the votaries, however, does not account in this case for the generally limited presence of male deities compared to that of goddesses. As already mentioned, votive offerings took many different forms, but it would be wrong to think that no matter what could be dedicated. Votive gifts were always connected to the identity of the dedicators as these offerings reflected in their form the particular occupation or preoccupation of those who made them60. Moreover, the amalgam of goddesses represented at RS do share certain character- istics: they are all deities concerned with female fertility, womanhood, and the nursing and upbringing of children. Such dedications to kourotrophic divinities who nurtured the young, indicate a concern for female reproductive processes and the physical development of infants61. Rather than the coincidence of availability, it was the concerns these gods addressed that caused their presence in this extra-urban location, an observation that brings us to the people worshipping them.

Social groups/worshippers Since none of the votives found at RS carry an inscription, except for two figurines signed by the coroplast Apelles, the identity of their donors is not known. Having said that, votive gifts, like other artefacts, carry the potential to reveal aspects of the identity of

55 Talloen 2015: 186-187. 56 Agelidis 2009; Ateş 2014. 57 Talloen 2015: 236-241. 58 Alroth 1989. 59 Alroth 1989: 65. 60 Osborne 1987: 185. 61 Cole 2004: 213-214. A PLACE ON THE FRINGE OF SAGALASSOS 279 the votaries, such as gender and status62. Especially when not simply reproducing modern prejudices about the significance of individual artefacts and their links with different categories of people, so that weapons, for example, have exclusively male associations and ornaments are associated with women, but considering their complete archaeological context, possible identity-related objects can be revealed63. Natural sanctuaries such as RS are normally located in the rural sphere and have therefore often been associated with simplicity (or even poverty) and worshippers from the lower strata of the population, something which has recently been criticised64. Natural sanctuaries are now understood as expressions of religious structures and needs, not as reflections of social hierarchies. Given its close proximity to the urban centre, RS was undoubtedly frequented by the people of Sagalassos. This is also corroborated by the finds. Not only were the tableware used for ritual feasting, as well as the votive offerings in the shape of figurines and plaques locally produced, but also the import of some of the votive gifts, including lead-glazed pottery and coins from Perge in Pamphylia, sea shells from the Mediterranean, and even millefiori glass vessels from possibly as far away as Egypt, indicate an urban rather than a rural origin for the sanctuary’s clientele. The activities of the urban populace were often reflected in the sanctuaries found immediately outside the town. These suburban sanctuaries acquired much of their character by being adopted by one or another social group. Some of these groups were part of official, civic groups, but not essentially so. The sanctuaries outside the city walls provided religious foci which were not constrained by the political space within the town. These cult sites provided the space in which a member of the community could be social without being political65. They are locations where aspects of religious practices other than the civic ones of polis religion, that have often dominated research into ancient religious life66, can be approached. RS provides a case where we can study religious practice of social groups outside of the dominant sphere of the polis. Even if associating material culture with social groups is a difficult issue, the subject matter of the terracotta figurines from RS already points to certain groups. Images of human individuals retrieved from RS consisted for the most part of busts of girls and women (79 % of all mortals), with boys and young males (6.5 %) a distant second. Other frequent types were female musicians, girl and boy students, and adolescents on horseback; adult men do not feature at all. As mentioned above, they can probably be identified as representations of the worshippers themselves or of the people for whom they were asking protection, being girls, boys and women. Figurines in the past were perceived as being imbued with personhood and when representing sexed bodies could be used in the forma- tion of gender and age67. They were therefore an appropriate item in rituals linked with the female life cycle.

62 Allison 2015. 63 Bradley 2000: 55. 64 Baumer 2014. 65 Osborne 1987: 168. 66 Kindt 2009. 67 Insoll 2017: 6. 280 P. TALLOEN ET AL.

As far as representations of deities are concerned, Aphrodite clearly stands out. She was the standard for female beauty and patron of the sphere of sexuality, and her cult was emphatically the preserve of women68. Nearly all other represented divinities – Eros, Tyche, Athena, Kybele, Isis and Hermes – are equally concerned with femininity or the vitality of children and the future of the family. In addition to the overriding female character of the figurines, other categories of votive offerings include objects typically associated with women (Fig. 20). For example, the tools used in spinning and weaving – essential skills of housewives and brides-to-be – were dedicated: needles, loom weights (Fig. 20a), spindles, spindle whorls, and distaffs. The distaff in particular was a tool usually associated with the status of women as matron or lady of the house (Fig. 20b)69. Other find categories such as (miniature) mirrors (Fig. 20d), worked bone combs and hair pins (Fig. 20c), ceramic and glass perfume bottles, and metal and glass pieces of jewellery (Fig. 20e-f), all belong to the sphere of beautification and are traditionally identified as archetypical female attributes70. The presence of copper-alloy writing utensils or styli among these finds may at first seem out of place (Fig. 21). Yet, when also considering the presence of terracotta figurines of seated children – boys and girls – holding a diptych or writing tablet on their lap, it is clear that literacy, as part of the education of children, was also one of the concerns that was addressed at the sanctuary. Overall then, the assemblage yields a consistent picture of womanhood as the dominant theme of the sanctuary, at least for the Roman Imperial period, and of women as the social group most probably responsible for their dedication. Given that these offerings occurred as part of small group celebrations of private social events as indicated by the ritual meals, their ritual efforts at RS most probably focused on the female life cycle with human fertility and reproduction, as well as the raising and education of children as central themes.

Abandonment of the cult site The accumulation of votive offerings appears to have come to an end during the early 3rd century CE, which would suggest that RS was abandoned as a cult site shortly after- wards. It is true that there are no signs for the production and use of terracotta figurines at Sagalassos after the first half of the 3rd century CE, at RS or elsewhere in the city until the 5th century CE71. At first sight, this could explain the absence of votive figurines, yet the same is also true for all other categories of votive gifts. Although the possibility of a worship leaving no material traces cannot be fully excluded, the nature of ancient religious practice in the region, focussing on the performance of rituals defined by symbols and objects72, makes this highly unlikely.

68 Pirenne-Delforge 2007. 69 Cremer 1996; Trinkl 2004. 70 Swift 2011; Allison 2015. 71 Talloen in press. 72 Talloen 2015: 7-11. A PLACE ON THE FRINGE OF SAGALASSOS 281

a b

Fig. 20. Overview of gendered votive offerings: c d a. terracotta loom weight; b. worked bone distaff; c. worked bone hair pin; d. fragment of the lead frame of a miniature mirror; e. copper-alloy pendant; f. glass bead (© Sagalassos Archaeologi- cal Research Project). e f

The reason for its abandonment is not clear, but the end date cannot be related to the rise of Christianity in the city as this phenomenon only made its mark from the later 4th-early 5th century CE onwards73. Internal developments within traditional polytheistic religion seem more likely. As already pointed out by Beatrice Caseau, the ancient religious landscape was a dynamic one: cults flourished and disappeared74. Christianity was not responsible for all those changes. There was an internal dynamic of the polytheistic cults as well, and this may have been responsible for the situation at RS. The cave sanctuary at Pergamon, for example, a site dedicated to the cult of the Mother Goddess, lost its popularity at the end of the 1st century BCE. The crockery of the site, as well as some of the votive offerings were then ritually buried inside the sanctuary75. A loss of popularity of the deities worshipped at the site seems unlikely at the beginning of the 3rd century CE. As goddess of love and beauty, Aphrodite – the main subject of worship at RS – undoubtedly continued to play a dominant role in the life of women at Sagalassos, given the reproductive concerns and family responsibilities she addressed. Moreover, she even achieved a public profile during the Middle Roman Imperial period (2nd-3rd centuries CE). This is suggested by the erection of monumental statues of the goddess in public spaces such as the Hadrianic Nymphaeum and the Roman Baths76, as well as the issue of a civic coinage type carrying the effigy of Aphrodite anadyomene accompanied by Eros during the reign of Severus Alexander (222-235 CE)77. While she was completely absent from manifestations of officially sanctioned worship during the Early Roman Imperial period, the goddess became prominent in the city-scape during the Mid- dle Roman Imperial period. The same is true for the Egyptian triad of Isis, Harpokrates and Sarapis. These foreign deities were no members of the civic pantheon when they entered the Pisidian cities through the activity of private individuals rather than as a result of any official intervention. This seems again to be confirmed by the earlier dates for the popular manifestations of their worship – like the 1st century CE figurines from RS – when compared with the public sources appearing from the later 2nd century CE onwards78. These kinds of developments are illustrative of an ongoing recuperation effort on the part of the civic officials whereby the ancestral religious tradition gradually enlarged itself to include originally private cults. The officials thus reduced the ‘tension’ between the official pantheon and the popular cults in order to prevent the former from becoming fossilised and to keep it relevant for the members of the community79. It is tempting to relate the abandonment of RS, focussed on the cult of several deities that obviously continued to remain popular in the religious life of the community of Sagalassos, to such an ‘officialisation’ of the cult(s). Such a process would have rendered the popular cult site at RS redundant as it was then replaced by an official cult site in the urban centre. It would

73 Talloen 2019b. 74 Caseau 1999. 75 Pirson et al. 2015; Engels 2019. 76 For the statue of Aphrodite from the Hadrianic Nymphaeum: Mägele et al. 2007: 485-486; for that of the Roman Baths: Waelkens et al. 2011: 10. 77 Levante and Weiss 1994: n° 1811. 78 Talloen 2015: 197. 79 Talloen 2015: 197. A PLACE ON THE FRINGE OF SAGALASSOS 283 also help to explain the total lack of monumentalisation at the site, in a period when several other cave sites in the region received a monumental make-over (see above). Whatever the case, it is clear that by the middle of the 3rd century CE cult activity at RS had come to a halt. Activity at the site is again attested during the (later) 4th century CE, in the form of shards of few vessels of local manufacture – which include early versions of (drinking) cup types 1A140-3 – present in nearly all the in situ deposits that are characterised by the mixed assemblages of offerings. It was already mentioned that the majority of these objects were very fragmented and weathered indicating that they had been exposed to the elements for quite some time before they were buried inside Rooms 2 and 3. As these deposits also included fragments of roof tiles, we appear to be dealing with a general clean-up of the sanctuary which caused the cavities to be filled up with everything that was found lying around, votives as well as roof tiles belonging to shelters built around the outcrop. So rather than the in situ deposition of votive offerings or the gradual accumulation of ritual waste – representing the regular tidying of the remains of ritual activities – the deposits were primarily the result of a single operation sometime in the 4th century CE. At this time, an involvement of Christianity is possible as Christians were becoming increasingly militant in the urban centre of Sagalassos during the second half of the 4th century CE; their actions resulted in the conversion or closure of several sanctuaries by the end of that century80. Likewise, they may very well have chosen to clean up the abundant remains of pagan worship at the extra-mural site of RS once and for all, something that may also help to explain the generally very fragmented nature of the votive offerings. It was on top of the clean-up deposits and contemporary floor levels throughout the spaces, that large blocks of limestone were found that can most probably be attributed to the collapsed ceiling of the complex. This suggests that after its definite abandonment in the late 4th century CE, the site was probably struck by the earthquake that also devastated the urban centre of Sagalassos around the middle of the 7th century CE. The seismic catastrophe caused some of the old connections between the spaces to be closed-off, making the central space, Room 4, largely inaccessible. It was followed by the deposition of colluvium in some spaces.

A Middle Byzantine place of refuge? Centuries later, a new floor level was arranged inside the now much diminished space of Room 1 on top of the partly collapsed ceiling and the colluvium that had accumulated, during the Middle Byzantine period. A total of 28 fragments representing a small number of vessels of Middle Byzantine date (late 12th-early 13th century) were found in floor deposits as well as in the fill of a pit containing some charcoal, ashes and faunal remains that can probably be related to food preparation (Fig. 22). In addition to shards of storage and cooking vessels (including a lid), the pottery included a number of fragments of a glazed plate in incised sgraffito ware, a relatively precious, imported table ware, only attested in significant numbers at Sagalassos in the short-lived Byzantine fort on top of Alexander’s

80 Talloen 2019b: 178-180. 284 P. TALLOEN ET AL.

Fig. 22. Fragments of Middle Byzantine pottery, including fragments of one yellow-glazed dish (right), a lid (centre bottom) and a handle of a cooking vessel (bottom left). The fragment middle left could be Modern and of local manufacture (© Sagalassos Archaeological Research Project).

Hill in the southern extremity of Sagalassos81. What exactly these later vessels represent at RS remains speculative. The latter half of the Middle Byzantine period (11th-12th centuries) was a time of conflict between the Byzantines on the one hand and the Seljuks of the Sultanate of Rum on the other, with Pisidia as a frontier region that changed hands between the two82. It was after the reconquest of the area by the emperor Johannes II Komnenos around 1120 that the fortified stronghold must have been established on Alexander’s Hill. By that time, the city of Sagalassos had been largely abandoned and replaced by a kastron or fortified settlement centred on the promontory that previously housed the sanctuary of the imperial cult83. In the years following the disastrous defeat of the emperor Manuel Komnenos at Myri- okephalon in 1176, the Seljuks again seized control of the area, most probably laying waste to the fort on Alexander’s Hill and ending the settlement of the kastron84. These were uncertain times for sure, and given the less than comfortable circumstances provided by the largely collapsed spaces of RS, hardly appropriate for normal housing, one is tempted to identify the occupation of the cave-like site as a temporary shelter for some Byzantine people. A fragment of a possible lantern (or stove) with so-called Kerbschnitt decoration that presumably belongs to this Middle Byzantine assemblage may suitably represent this renewed and short-lived use of RS.

CONCLUSIONS

In spite of many years of illegal digging which destroyed large parts of the archaeological record, the rescue excavations at RS produced unexpectedly rich results, informing us about the historical use of the cave-like complex and the nature of the ancient activities

81 Vionis et al. 2010. 82 Belke and Mersich 1990: 102-124. 83 Poblome et al. 2017. 84 Vionis et al. 2010: 424-425. A PLACE ON THE FRINGE OF SAGALASSOS 285 taking place there. The archaeology of RS documents different functions throughout history of its usage – burial ground, sanctuary, shelter – in what remained essentially the same location. Indeed, notwithstanding its longevity and popularity, as evidenced by the enor- mous number of artefacts that was deposited at the site, no grandiose interventions could be ascertained there. This indicates that while the intrinsic qualities of the space undoubtedly played an important role, they cannot be considered deterministic for the role it played in the human landscape. The natural landform was given a cultural significance through the activities that took place there, as reflected by the material residue of these actions. After a brief use as a burial site during the Late Chalcolithic period, people returned to this location after the city-state of Sagalassos came into existence during the Hellenistic period. It was the combination of objects as a whole and the very context in which these were used and placed that made it possible to identify the site as a sanctuary, more particularly, a site of popular worship. RS thus offered a unique glimpse into an aspect of ancient life not previously known from Sagalassos and the possibility to understand religious cus- toms of one or more social groups other than the dominant elite represented at the shrines and temples of the city centre. The detritus of the rituals that took place there – ritual dining and votive deposition – was used to describe the social groups involved. The picture that emerged from the pottery was that of a sanctuary that presumably was visited by small groups of people who had close (family) ties. While the lack of written sources at RS will never allow us to obtain complete certainty concerning the composition of the participating group, representative votive offerings in the form of terracotta figurines leave no doubt that female concerns stood at the heart of the cult practiced there. Issues of love, sexuality, motherhood, childhood, education, and health, were all clearly brought to bear at the sanctuary. The presence of other categories of gendered material culture such as textile tools and elements of personal adornment, provides corroborating evidence to that extent. Con- sequently, it is permissible to assume that these gifts were primarily, if not exclusively, dedicated by (young) women, reflecting their (future) social roles as wives and mothers. Interestingly, not the Christianisation of the city but local religious developments must have been responsible for the eventual demise of RS already in the 3rd century CE, empha- sising once again the dynamic nature of ancient religious practice in spite of the conservative character that is generally attributed to it. Christianity, however, does appear to have been responsible for erasing all traces of cult at the site at the end of the 4th century. After a seismic catastrophe, probably during the 7th century, made the covered rock crevices largely inaccessible, the site only served as a temporary refuge in Middle Byzantine times before being abandoned.

REFERENCES

AGELIDIS, S. 2009 Cult and Landscape at Pergamon. In: C. Gates, J. Morin and Th. Zimmermann (eds.), Sacred Landscapes in Anatolia and Neighboring Regions (BAR International Series 2034). Oxford: Archaeopress, 47-54. 286 P. TALLOEN ET AL.

ALCOCK, S.E. and J. REMPEL 2006 The More Unusual Dots on the Map. ‘Special-Purpose’ Sites and the Texture of Landscape. In: P. Guldager Bilde and V. Stolba (eds.), Surveying the Greek Chora. The Black Sea Region in Comparative Perspective (Black Sea Studies 4). Aarhus: Aarhus University Press, 27-46. ALLISON, P.M. 2015 Characterizing Roman Artefacts to Investigate Gendered Practices in Contexts without Sexed Bodies. American Journal of Archaeology 119(1): 103-123. ALROTH, B. 1989 Greek Gods and Figurines. Aspects of the Anthropomorphic Dedications. Uppsala: Academia Upsaliensis. ATEŞ, G. 2014 Pergamon’da Doğa ve Kült. Ana Tanrıça İnancı ve Doğal Kutsal Alanlar / Nature and Cult in Pergamon. Meter Worship and Natural Sanctuaries. In: F. Pirson and A. Scholl (eds.), Pergamon. Anadolu’da Hellenistik Bir Başkent / Pergamon. A Hellenistic Capital in Anatolia. Istanbul: Yapı Kredi Yayınları, 422-437. BARFOED, S. 2018 The Use of Miniature Pottery in Archaic-Hellenistic Greek Sanctuaries. Considera- tions on Terminology and Ritual Practice. Opuscula 11: 111-126. BAUMER, L. 2014 Où le paysan faisait une pause pour offrir quelque modeste don – Les sanctuaires ruraux en Grèce. Entre pauvreté romantique et réalité archéologique. In: E. Galbois and S. Rougierblanc (eds.), La pauvreté en Grèce ancienne. Formes, représentations, enjeux (Scripta Antiqua 57). Paris: Ausonius - De Boccard, 97-104. BELKE, K. and N. MERSICH 1990 Phrygien und Pisidien (Tabula Imperii Byzantini 7). Vienna: Verlag der Österrei- chischen Akademie der Wissenschaften. BENDLIN, A. 2007 Purity and Pollution. In: D. Ogden (ed.), A Companion to Greek Religion. Oxford: Blackwell, 178-189. BRADLEY, R. 2000 An Archaeology of Natural Places. London: Routledge. BUMKE, H. 2015 Griechische Gärten im sakralen Kontext. In: K. Sporn, S. Ladstätter and M. Kerschner (eds.), Natur, Kult, Raum, Akten des internationalen Kolloquiums Paris-London- Universität Salzburg, 20.-22. Jänner 2012 (Österreichisches Archäologisches Institut Sonderschriften 51). Vienna: Österreichisches Archäologisches Institut, 45-61. CASEAU, B. 1999 Sacred landscapes. In: G.W. Bowersock, P. Brown and O. Grabar (eds.), Late Antiquity. A Guide to the Postclassical World. Cambridge (Mass.): Harvard University Press, 21-59. COLE, S.G. 2004 Landscapes, Gender and Ritual Space. The Ancient Greek Experience. Berkeley: University of California Press. CREMER, M. 1996 Venuskunkeln aus Kleinasien. Archäologischer Anzeiger: 135-144. DAEMS, D., M. VAN DER ENDEN, P. TALLOEN and J. POBLOME 2019 The Hellenistic Pottery Repertoire made at Sagalassos, SW Anatolia. In: A. Peignard-Giros (ed.), Daily Life in a Cosmopolitan World: Pottery and Culture A PLACE ON THE FRINGE OF SAGALASSOS 287

during the Hellenistic Period. Proceedings of the 2nd Conference of IARPotHP, Lyon, November 2015, 5th-8th (IARPotHP 2). Vienna: Phoibos, 81-96. DIGNAS, B. 2017 A Day in the Life of a Greek Sanctuary. In: D. Ogden (ed.), A Companion to Greek Religion. Oxford: Blackwell, 163-177. DURAND, J.-L. and A. SCHNAPP 1989 Sacrificial slaughter and initiatory hunt. In: C. Berard (ed.), A City of Images: Iconography and Society in Ancient Greece. Princeton: Princeton University Press, 52-70. DURU, R. 2008 MÖ 8000’den MÖ 2000’e: Burdur-Antalya Bölgesi’nin Altıbin Yılı. Antalya: Suna-İnan Kıraç Akdeniz Medeniyetleri Araştırma Enstitüsü. EKINCI, H.A. and Y. ZENGER 2017 İnarası Mağarası Kutsal Alanı Kurtarma Kazısı 2016. Müze Kurtarma Kazıları Sem- pozyumu 26: 131-142. ENGELS, B. 2015 Festbankette im Grottenheiligtum von Pergamon. In: J.-A. Dickmann and A. Heinemann (eds.), Vom Trinken und Bechern: Das antike Gelage im Umbruch. Katalog zur gleichnamigen Ausstellung Archäologische Sammlung der Universi- tät Freiburg 26. April-28. Juni 2015. Freiburg: Universitätsdruckerei Freiburg, 91-99. 2019 Kultpraxis, Akteure und Atmosphäre in pergamenischen Grottenheiligtums des 2. und 1. Jahrhunderts vor Christus. In: B. Engels, S. Huy and C. Steitler (eds.), Natur und Kult in Anatolien (Byzas 24). Istanbul: Ege Yayınları, 117-143. HARMANŞAH, Ö. 2014 Introduction. Towards an Archaeology of Place. In: Ö. Harmanşah (ed.), Of Rocks and Water. Towards an Archaeology of Place. Oxford: Oxbow Books, 1-12. HUYSECOM-HAXHI, S. and A. MULLER 2007 Déesses et/ou mortelles dans la plastique de terre cuite. Réponses actuelles à une question ancienne. Pallas 75: 231-247. 2015 Figurines grecques en contexte: Présence muette dans le sanctuaire, la tombe et la maison. Villeneuve d’Asq: Septentrion. INSOLL, T. 2017 Miniature Possibilities: An Introduction to the Varied Dimensions of Figurine Research. In: T. Insoll (ed.), Oxford handbook of Prehistoric Figurines. Oxford: Oxford University Press, 3-15. KARTAL, M. and C.M. EREK 1998 The burials in Öküzini cave (SW Anatolia). In: M. Otte (ed.), Préhistoire d’Anato- lie: Genèse de deux mondes. Liège: Université de Liège, 551-558. KINDT, J. 2009 Polis Religion – A Critical Appreciation. Kernos 22: 9-34. LEMAÎTRE, S., S.Y. WAKSMAN, M.-C. ARQUÉ, E. PELLEGRINO, C. ROCHÉRON and B. YENER- MARKSTEINER 2013 Identité régionale et spécificités locales en Lycie antique: l’apport des céramiques culinaires. In: P. Brun, L. Cavalier, K. Konuk and F. Prost (eds.), Euploia. La Lycie et la Carie antiques. Dynamiques des territoires, échanges et identités. Actes du colloque de Bordeaux, 5, 6 et 7 Novembre 2009 (Mémoires d’Ausonius 34). Paris: Editions de Boccard, 189-212. 288 P. TALLOEN ET AL.

LEVANTE, E. and P. W EISS (eds.) 1994 Sylloge nummorum graecorum. France, 3. Cabinet des Médailles. Paris: Biblio- thèque nationale. MARTIN, S.R. and S.M. LANGIN-HOOPER 2018 In/Complete. An Introduction to the Theories of Miniaturization and Fragmenta- tion. In: S.R. Martin and S.M. Langin-Hooper (eds.), The Tiny and the Frag- mented: Miniature, Broken, or Otherwise Incomplete Objects in the Ancient World. Oxford: Oxford University Press, 1-23. MAVRIDIS, F., J.T. TENSEN and L. KORMAOPOULOU 2013 Introduction. In: F. Mavridis and J.T. Jensen (eds.), Stable Places and Changing Perceptions. Cave Archaeology in Greece (BAR International Series 2558). Oxford: Archaeopress, 1-16. MÄGELE, S., J. RICHARD and M. WAELKENS 2007 A Late-Hadrianic Nymphaeum at Sagalassos (Pisidia, Turkey): A Preliminary Report. Istanbuler Mitteilungen 57: 469-504. MIKALSON, J.D. 2004 Ancient Greek Religion (Blackwell Ancient Religions). Oxford: Blackwell. MYLONOPOULOS , J. 2006 Greek Sanctuaries as Places of Communication through Rituals. An Archaeological Perspective. In: E. Stavrianopouloou (ed.), Ritual and Communication in the Graeco- Roman World (Kernos Supplément 16). Liège: Centre international d’étude de la religion grecque antique, 69-110. 2008 Natur als Heiligtum – Natur im Heiligtum. Archiv für Religionsgeschichte 10: 51-83. NOHLEN, K. and W. RADT 1978 Kapıkaya, ein Felsheiligtum bei Pergamon (Altertümer von Pergamon 12). Berlin: De Gruyter. OSBORNE, R. 1987 Classical Landscape with Figures. The Ancient Greek City and Its Countryside. London: George Philip. 2004 Hoards, Votives, Offerings. The Archaeology of the Dedicated Object. World Archaeology 36(1): 1-10. PERRIN, F. 1991 Le site de la Chuire (camp de Larina, Hières-sur-Amby, Isère) et l’Isle Crémieu à l’Age du Fer. In: A. Duval (ed.), Les Alpes à l’Age du Fer. Paris: CNRS Editions, 21-48. PIRENNE-DELFORGE, V. 2007 Something to Do with Aphrodite: Ta Aphrodisia and the Sacred. In: D. Ogden (ed.), A Companion to Greek Religion. Oxford: Blackwell, 311-323. PIRSON, F., G. ATEȘ and B. ENGELS 2015 Die neu entdeckten Felsheiligtümer am Osthang von Pergamon – ein innerstädti- sches Kultzentrum für Meter-Kybele? In: K. Sporn, S. Ladstätter and M. Kerschner (eds.), Natur, Kult, Raum, Akten des internationalen Kolloquiums Paris-London- Universität Salzburg, 20.-22. Jänner 2012 (Österreichisches Archäologisches Insti- tut Sonderschriften 51). Vienna: Österreichisches Archäologisches Institut, 281- 301. POBLOME, J. 1999 Sagalassos Red Slip Ware. Typology and Chronology (Studies in Eastern Mediter- ranean Archaeology 2). Turnhout: Brepols Publishers. A PLACE ON THE FRINGE OF SAGALASSOS 289

POBLOME, J., TALLOEN, P. and E. KAPTIJN 2017 Sagalassos. In: P. Niewöhner (ed.), The Archaeology of Byzantine Anatolia. From the End of Late Antiquity until the Coming of the Turks. Oxford: Oxford Univer- sity Press, 302-311. POBLOME, J. et al. 2018 Sagalassos’ta 2016 Sezonu Kazılar, Restorasyon Çalışmaları ve Malzeme Etütleri. In: 39. Kazı Sonuçları Toplantısı, Cilt 3. Ankara, 211-238. 2019 Sagalassos’ta 2017 Sezonu Kazılar, Restorasyon Çalışmaları ve Malzeme Etütleri, In: 40. Kazı Sonuçları Toplantısı, Cilt 2. Ankara, 485-506. ROTROFF, S.I. 1991 Attic West Slope Vase Painting. Hesperia 60(1): 59-102. ROWAN, Y.M. 2018 The spiritual and social landscape during the Chalcolithic period. In: A. Yasur- Landau, E.H. Cline and Y.M. Rowan (eds.), The Social Archaeology of the Levant. Cambridge: Cambridge University Press, 122-145. SWIFT, E. 2011 Personal Ornament. In: L. Allason-Jones (ed.), Artefacts in Roman Britain. Their Purpose and Use. Cambridge: Cambridge University Press, 194-218. TALLOEN, P. 2015 Cult in Pisidia. Religious Practice in Southwestern Asia Minor from Alexander the Great to the Rise of Christianity (Studies in Eastern Mediterranean Archaeology 10). Turnhout: Brepols Publishers. 2019a The Rock Sanctuary: Nature, cult and marginality in the periphery of Sagalassos. In: B. Engels, S. Huy and C. Steitler (eds.), Natur und Kult in Anatolien (Byzas 24). Istanbul: Ege Yayınları, 177-197. 2019b The rise of Christianity at Sagalassus. In: S. Mitchell and P. Pilhofer (eds.), Early Christianity in Asia Minor and Cyprus: From the Margins to the Mainstream (Early Christianity in Asia Minor 3). Leiden: Brill, 164-201. 2020 Introducing the smaller brother: the production of terracotta figurines at Sagalassos. In: In: A. Mörel et al. (eds.), Production, Trade and Economy in Pisidia and its Surrounding Regions: International Symposium Proceedings (Pisidian Studies 2). Isparta: Isparta University Press, 166-181. TALLOEN, P. and S. ÖZDEN-GERÇEKER 2020 Out of the Rock: Terracotta Figurines from the Rock Sanctuary at Sagalassos kept at the Sadberk Hanım Museum in Istanbul. Les Carnets de l’ACoSt 21. URL: http:// journals.openedition.org/acost/2601 TALLOEN, P. and J. POBLOME 2016 The 2014 and 2015 Control Excavations on and around the Upper Agora of Saga- lassos. The Structural Remains and General Phasing. Anatolica 42: 111-150. TALLOEN, P., J. POBLOME and H.A. EKINCI 2015 Salvage Excavations at a Rock Sanctuary Near Sagalassos in 2014 / Sagalassos Yakın- larındaki Kaya Kült Alanında Kurtarma Kazıları 2014. Anmed 13: 108-113. TALLOEN, P. and M. WAELKENS 2004 Apollo and the Emperors, I. The Material Evidence for the Imperial Cult at Saga- lassos. Ancient Society 34: 171-216. THEODOROPOULOU, T. 2013 The sea in the temple? Shells, fish and corals from the sanctuary of the ancient town of Kythnos and other marine stories of cult. In: G. Ekroth and J. Wallensten (eds.), 290 P. TALLOEN ET AL.

Bones, Behaviour and Belief: The Zooarchaeological Evidence as a Source for Ritual Practice in Ancient Greece and Beyond. Stockholm: Svenska Institute i Athen, 197- 222. TRINKL, E. 2004 Zum Wirkungskreis einer kleinasiatischen Matrona anhand ausgewählter Funde aus dem Hanghaus 2 in Ephesos. Jahreshefte des Österreichischen Archäologischen Institutes 73: 281-303. VANDAM, R. et al. 2017 Living on the margins. First results from the Dereköy Archaeological Survey of the Sagalassos Project in the Western Taurus Mountains. In: S. Steadman and G. McMahon (eds.), The Archaeology of Anatolia Vol. II: Recent Discoveries (2015-2016). Newcastle: Cambridge Scholars Publishing, 321-346. VAN DER ENDEN, M., J. POBLOME and P.M. BES 2014 From Hellenistic to Roman Imperial in Pisidian Tableware: The Genesis of Sagal- assos Red Slip Ware. In: H. Meyza (ed.), Late Hellenistic to Mediaeval Fine Wares of the Aegean Coast of Anatolia. Their Production, Imitation and Use. Krakow: Archeobooks, 81-93. 2018 Sagalassian Mastoi in an Eastern Mediterranean Context. In: S. Drougou (ed.), 9th Scientific Meeting on Hellenistic Pottery, Thessaloniki December 5-9th 2012. Athens, 925-945. VILLING, A. 2017 Don’t Kill the Goose That Lays the Golden Egg? Some Thoughts on Bird Sacrifices in Ancient Greece. In: S. Hitch and I. Rutherford (eds.), Animal Sacrifice in the Ancient Greek World. Cambridge: Cambridge University Press, 63-102. VIONIS, A. et al. 2010 A Middle-Late Byzantine pottery assemblage from Sagalassos: Typo-chronology and socio-cultural interpretation. Hesperia 79(3): 423-464. WAELKENS, M. et al. 2011 Sagalassos. City of Dreams. Catalogue. Gent: OKV.